Lessons in Estimating Agile vs. Waterfall Agile and Waterfall. Jerry Richardson, PMP Sohail Thaker, PMP

Lessons in Estimating

Agile vs. Waterfall

Agile and Waterfall

2

• Why do we estimate?

• Developing Waterfall estimates • Developing Agile estimates

• When to use each technique

Workshop Overview

• When to use each technique • Keys to success

4

Estimating vs. Planning

Develop Project Framework Determine Resource Requirements Prepare a Budget Plan a Plan a Project Project Develop Work Plan Develop a Schedule Develop Project Controls • Tasks

Why Develop Work Plan?

• Funding

• Resources • Expectations

It’s management’s job to vet our plan! “Realistic & Defensible”

6

Why Develop Work Plan?

• Who needs to be involved? • What will each person do?

What Do We Estimate?

• Work Effort

• Elapsed Time • Cost

8

Levels of Estimates

Project level estimate

Construct Deploy Design Analyze Top Down Construct Design Analyze

Bottom Up Bottom Up Bottom Up

Phase level estimates

10

Bottom Up Estimating Technique

WBS - Example

PREPARE FOR TRIP MAKE TRAVEL ARRANG’TS PACK MAKE “HOUSE” ARRANG’TS TRAVEL TO AIRPORT COLLECT CANCEL ARRANGE CONFIRM MAKE RESERV’NS COLLECT TRAVEL DOCUMENTS CANCEL NEWSPAPER ARRANGE HOUSE CHECK PREPARE FIND SUITCASE PACK CLOTHES DECIDE WHAT TO TAKE WASH CLOTHES IRON CLOTHES CONFIRM DATES 12

Bottom Up Estimating Technique

1. Break job into tasks (WBS) 2. Develop a task list

Estimating Spreadsheet

Task Deliverable

Confirm Dates Dates

Confirmed Make Reservations Reserv’ns Collect Travel Docs Passport, E-ticket …(Etc.) 14

Bottom Up Estimating Technique

1. Break job into tasks (WBS) 2. Develop a task list

3. For each task, estimate:

S – shortest effort S – shortest effort L – longest effort

M – most likely effort

Estimating Spreadsheet

Task Deliverable Assumptions S L M

Confirm Dates Dates

Confirmed

Boss in office 5 20 5 Make

Reservations

Res’ns In-house travel agent 15 30 20 Collect Travel Docs Passport, E-ticket …(Etc.) 16

Bottom Up Estimating Technique

4. Calculate the best estimate, E

E = S + L + 4M 6

Estimating Spreadsheet

Task Deliverable Assumptions S L M E

Confirm Dates Dates

Confirmed

Boss in office 5 15 5 7 Make

Reservations

Res’ns In-house travel agent 15 30 20 21 Collect Travel Docs Passport, E-ticket …(Etc.) 18

Bottom Up Estimating

4. Calculate the best estimate, E

E = S + L + 4M 6

5. Add contingency (10-20%) 5. Add contingency (10-20%)

Contingency Factors

• Project size • Project duration • Project organization • Technical complexity • Technical complexity

• Stability of business area • Project team experience

20

Levels of Estimates

Project level estimate

Construct Deploy Design Analyze Top Down Construct Design Analyze

Bottom Up Bottom Up Bottom Up

Top Down Estimating

• All Top Down estimates are a range • Techniques

– Equation

• % of effort for each phase

– Comparison

• similarity to another company project

– Analogy

• similarity to a database of projects

22

Top Down Estimating

Equation Approach – “Ball Park” Example

% of Total Time Initiate Initiate Analyze 5 - 15% Design 15 - 30%

Top Down Estimating

Equation Approach – “Ball Park” Example

Initiate 2 days

Analyze (est’d) 8 days

10 days 10 days

5% of total project = 10 days total project = 200 days 15% of total project = 10 days total project = 67 days

“Ball Park” Estimate = 70 – 200 days

24

Keys to Success

• Start with a “ball park” and compare to your bottom up estimate

• Use your team • Ask for input

• Estimate next phase as part of current phase

• Agile Manifesto • Many versions

• Focused on software development • Incorporates many best practices

What is Agile?

• Incorporates many best practices • Focus on “features”

• Business needs to be very involved • Iterative delivery of features

26

Construct Deploy Design Analyze WATERFALL AGILE

Agile Iterations

AGILE

• Developing Agile Estimates

• What is Velocity and how do you use it

Agile Estimating

• What is Velocity and how do you use it for Planning & Refining Estimates?

28

Law of Diminishing Returns

• The more effort we put into something, the better result. Right?

• Law of Diminishing Returns – additional estimation effort yields very little value beyond a certain point.

Estimating: Accuracy vs. Effort

30

Agile Assumptions

• Estimates are shared • Core team is dedicated • Complete a feature/user

story in one iteration

• Final product details are • Final product details are

not completely known • Business support

User Stories, Epics & Themes

32

Point Estimating Scales

• Fibonacci Sequence – 1,2,3,5,8 – (+ 13,20,40,100 for epics, themes) • T-Shirt Sizing • T-Shirt Sizing

– Small, Med, Large, X-Large

• 1,2,4,8

34

Points Example

• Feature 1 - 8 points (M) • Feature 2 - 8 points (M) • Feature 3 - 3 points (S) • Feature 4 - 13 points (L) • Feature 5 - 3 points (S) • Feature 6 - 40 points (XL) ... • ETC - NN points _______________________________ TOTAL = 500 points for project

36

Validating Story Point Estimates

• Time Estimates

– Pick some sample user stories – Break down to task level

– Estimate hours – Estimate hours

– Calculate how long it will take, in real hours, to develop a small, a medium, etc.

Agile Estimating: Results

• Master Story List with points for each story • Apply the real estimates to get an overall

estimate of the size of the project • Build in some contingency

• Build in some contingency

• Relatively accurate estimates for next few iterations

• Minimized time estimating

38

What is Velocity?

• Velocity determination • Self-correcting

Watch for Velocity Trends

• Example:

– Estimated 500 points for a project

– Team initially estimates they can complete 50 points per iteration

– Initial estimate = 500/50 = 10 iterations to – Initial estimate = 500/50 = 10 iterations to

complete project

– If first iterations average at 75 points, then team was too cautious and 500/75 = 7

iterations

– If first iterations average at 40 points, then team was too optimistic and 500/75 = 13 iterations

40

Track Progress Using Points

• Total Points Completed, Points Remaining

• Percent of Points completed

• Number of iterations to Complete

• Number of weeks/months to complete • Number of weeks/months to complete • Impact of changing resources to velocity • PM can add in factors once trends

Using Velocity to Plan

• Reduce scope

– deselect some specific stories (features)

• Increase capacity (size) of team

– add to team velocity

• Reduce cost by reducing resources • Reduce cost by reducing resources

– slow down velocity

42

PROS

• Fail Fast

• Working deliverables provided frequently

• Don’t need fully detailed requirements to start

• Tight feedback loop allows for requirement changes throughout the project

Agile: Pros & Cons

CONS

• No detailed design phase makes it difficult to

estimate the cost and duration of the whole project • Best applied to software development

44

PROS

• Detailed design phase gives better understanding to estimate the duration of the whole project

• Detailed design done by the experienced developers allow for more inexperienced developers (or

outsourcing of development)

• Enforced discipline by clear phasing with defined start and end points (milestones)

Waterfall: Pros & Cons

• Enforced discipline by clear phasing with defined start and end points (milestones)

CONS

• Big overhead caused by detailed early-phase planning, as plans tend to change

Choosing an Approach

46

Buy vs. Build (Waterfall, Agile, Blended) • Business Availability • Organization Maturity • Technical Complexity

How Do We Choose?

• Technical Complexity • Understood vs. New • Market Maturity • Competitive Advantage • Logical Feature Sets

Selection Criteria

Buy Build Waterfall Build Agile Don’t Do It Business Availability Very Low Low / Medium Medium / High Business Maturity Start-Up Organization Mature Organization 48

Selection Criteria

Buy Build Waterfall Build Agile Don’t Do It Technical Complexity Low Medium High Understood vs. New New Territory

Selection Criteria

Buy Build Waterfall Build Agile Don’t Do It Market Maturity No packages available Some packages available Many packages available

Competitive Advantage

Not required

Required

50

Selection Criteria

Buy Build Waterfall Build Agile Don’t Do It

Logical Feature Sets

Small discrete groups

Big – Need All

• Look at all the aspects before selecting the appropriate approach

• Both techniques have pros & cons

• Agile and Waterfall –

know when to use each technique

Conclusions

technique

• “Buy” when you can!

52

Some PM Truths Don’t Change…

• The more ridiculous the deadline the more

money will be wasted trying to meet it.

• The first 90% of a

project takes 90% of the time the last 10% takes the other 90%.

54

jrichardson @ethier.ca sthaker

Thanks!

sthaker @ethier.ca